Because of JPC’s hardware agnosticism, these stories apply equally well to non-x86 hardware,
the ideal solution for thin clients, and users get their favorite x86 environment in all cases.
Flexible Auditing and Support
The screen, keyboard, and mouse of a running JPC instance could be viewed and overridden
remotely by a suitably authorized system. Fraud could be effectively monitored by remotely
collecting keystrokes and screenshots taken of suspicious activity to form an evidence base.
Given the low-level hardware access made possible by JPC, this feature cannot be subverted
by the technically able person who attempts to detect and remove monitoring software from
within the emulated operating system. Even a user given administrator rights (on the guest
system) would not be able to escape the monitoring, no matter how knowledgeable.
An auditing JPC system could simply record activity, scanning and flagging actions in real time,
or go one step further and prevent certain actions. For example, in collaboration with suitable
server software, a monitored instance could scan the video output for inappropriate images
and then obscure them (or replace them with other content) at the virtual video card level.
Such low-level monitoring means users could not subvert content protection systems by
installing alternative viewing software.
Remote assistance could be far more effective if a help desk could literally see exactly what the
entire screen was doing, and directly interact with the keyboard and mouse at the virtual
hardware level. This would be possible even when JPC ran operating systems for which remote
access was never implemented—for example, DOS, whose use is ongoing in many industries
and countries around the world.
Flexible Computing Anywhere
Rather than run the main emulation on a local resource and merely import data from remote
resources as necessary, the core emulation could be carried out on a central “JPC” server.
Because JPC needs only a standard JVM to operate, this central JPC server could be based on
hardware that is completely different from a normal x86 PC. There are some candidates already
that could achieve this, and JPC has already been demonstrated on a 96 core Azul compute
appliance. Other possibilities include Sun’s Niagara-based servers and systems built from
mobile phone technology (JPC has already booted DOS on a Nokia N95, an ARM11-based
system).
But why centralize a server to run all these JPC instances? Presumably any resource on the
Internet could run a JPC instance on behalf of anyone else, with the screen output and user
input being piped via the network to the virtual machine owner. In this way of working, the
world is viewed as “N” users with “M” machines, and there is no fixed mapping of hardware
ownership relationships between users in the former group and machines in the latter. If a
machine is idle, any one of the users can use it, remotely launching a JPC instance to work on
230 CHAPTER NINE